Re: [cowiki-dev] The new coWiki parser
"John (yt) Hogenmiller" <[email protected]> Thu, 22 Sep 2005 22:50:44 -0400
| Newsgroups | gmane.comp.php.cowiki.devel |
|---|---|
| Message-ID | <[email protected]> |
> Also, as a point of note, we're not dealing with "black magic" but > "bad magic"... we should be nothing, if not consistent when referring > to ill defined areas of doubt and uncertainty. In my opinion, the bad > magic bug has been fixed in CVS HEAD since late July of this year. > John-- > > I really did mean Black Magic in the sense that no one (Daniel included) > seems to know how the parser works, or how to modify it (reliably). I > wasn't referrring to the data-loss bug. I think /that/ was fixed when > Daniel disabled Kai's auto-recover code. > > Paul > Ahh... well, in that case, I misunderstood you. I usually lump magic into "real" and "psychological", and in my mind, any "real" magic is by definition bad and dark. The reason I noticed it was because I did a search on cowiki dark magic and only got one message, whereas cowiki bad magic returned many a conversation piece. So, in this case, dark magic refers to the parser and bad magic is the data-loss bug. I think a few people know how the parser *works*, but reliably modifying pretty much the only thing in the project that isn't a plugin is in fact a major undertaking in any project. Recently, I have been thinking that it could be possible to /move/ the parser out from core and adding alternative parsers to the mix, choosing your parser in the configuration. However, once you have documents created, you couldn't just "switch" parsers and expect all of your documents to work as before (of course, if one parser is expected to behave exactly the same as the other, you could expect it). I don't want to experiment with that until after the next release. -- YourTech, LLC - http://yourtech.us/ (this account is used for mailing lists)